基于小波变换的整车驾驶性评价信号去噪

整车驾驶性评价试验采集的加速度信号中混有噪声,严重影响了数据的准确性。针对整车驾驶性评价试验中采集的加速度数据存在的噪声对驾驶性评价指标值的准确性产生影响的问题,提出一种适合于整车驾驶性评价试验数据的小波去噪方法:根据整车加速度数据特征,初步选择备选的小波基函数,通过评价信噪比和均方根误差确定最优的小波基函数和阈值选取规则组合,在此基础上对含噪信号进行多个尺度的分解,通过评价由平滑度和均方根误差构造的复合指标确定最优小波分解层数,从而实现对噪声信号的滤除。对一换挡工况的加速度试验数据采用上述方法的去噪并进行分析,分析结果表明,该小波去噪方法不仅较好地保留了换挡工况中用于评价驾驶性的振动与冲击指标特征,并且能够有效地提取信号的有用频率成分,保证了驾驶性评价指标值的准确性。

Signal De-noising Method for Whole Vehicle Drivability Evaluation Based onWavelet Transform

The acceleration collected by the drivability evaluation are mixed with noise, seriously affecting the accuracy of the data. Aiming at the problem that the accuracy of drivability evaluation indexes are affected by the noise in the data collected by the vehicle under different working conditions, a wavelet de-noising method is proposed for the pretreatment of the drivability evaluation data: The optimal wavelet basis function and threshold selection rules are determined by evaluating the signal-to-noise ratio and the root mean square error. The signal is decomposed by multiple scales. The wavelet decomposition layer is determined by the composite index constructed by the smoothness and root mean square error structure. So as to realize the filtering of the noise. The results of the analysis of the acceleration of the same gear shift condition show that the wavelet de-noising method not only retains the vibration and shock in the shift condition but also can effectively extract the useful frequency of the signal components and the evaluation index of the shift condition.